Hi Arnd,
On 09.05.2014 12:56, Arnd Bergmann wrote:
On Friday 11 April 2014, Daniel Lezcano wrote:
No more dependency on the arch code. The platform_data field is used to set the PM callback as the other cpuidle drivers.
Signed-off-by: Daniel Lezcano daniel.lezcano@linaro.org Reviewed-by: Viresh Kumar viresh.kumar@linaro.org Reviewed-by: Bartlomiej Zolnierkiewicz b.zolnierkie@samsung.com
This has just shown up in linux-next and broken randconfig builds.
diff --git a/arch/arm/mach-exynos/exynos.c b/arch/arm/mach-exynos/exynos.c index fe8dac8..d22f0e4 100644 --- a/arch/arm/mach-exynos/exynos.c +++ b/arch/arm/mach-exynos/exynos.c @@ -221,8 +221,9 @@ void exynos_restart(enum reboot_mode mode, const char *cmd) } static struct platform_device exynos_cpuidle = {
.name = "exynos_cpuidle",.id = -1,
.name = "exynos_cpuidle",.dev.platform_data = exynos_enter_aftr,.id = -1,};
This is wrong on many levels, can we please do this properly?
- The exynos_enter_aftr function is compiled conditionally, so you can't just reference it from generic code, or you get a link error.
+1
- 'static struct platform_device ...' has been deprecated for at least a decade, stop doing that. For any platform devices that get registered, there is platform_device_register_simple().
+0.5
The missing 0.5 is because you can't pass platform data using platform_device_register_simple(). There is platform_device_register_resndata(), though.
- There shouldn't need to be a platform_device to start with, this should all come from DT. We can't do this on arm64 anyway, so any code that may be shared between arm32 and arm64 should have proper abstractions.
-1
Exynos cpuidle is not a device on the SoC, so I don't think there is any way to represent it in DT. The only thing I could see this is matching root node with a central SoC driver that instantiates specific subdevices, such as cpufreq and cpuidle, but I don't see any available infrastructure for this.
Best regards, Tomasz